Non-Parametric Bayesian Techniques for Spatial Temporal Models, Optimisation and Decision Making

Abstract: The use of Bayesian techniques for modelling spatial temporal phenomena has extensively increased over the last decade, providing flexibility and uncertainty quantification for inference and prediction. This talk focuses on how to place Gaussian Process models over complex phenomena and explores how the information from these models can be used for flexible uncertainty aware decision making. This talk provides examples of the application and advantages of using these techniques for environmental monitoring, quantitative social sciences, criminology and human behaviour.

 

(Esta charla será dictada en inglés).

 

Date: Nov 02, 2017 at 14:30 h
Venue: Beauchef 851, Torre Norte, Piso 7. Sala de Seminario John Vob Neumann CMM.
Speaker: Roman Marchant
Affiliation: The University of Sydney
Coordinator: Prof. Felipe Tobar
Abstract:
PDF

Posted on Oct 30, 2017 in Seminario Aprendizaje de Máquinas, Seminars